CL-AI

A minimal command-line tool to get bash commands from local LLMs.

Features

• Native command-line interface that preserves your terminal, with an improved Text User Interface (TUI) powered by the rich library.
• Works both interactively and as a one-shot command.
• Color-coded and well-formatted responses.
• Command history and configuration accessible via interactive commands or command-line arguments.
• Loading spinner for feedback during LLM queries.
• Connects to local LLMs like Ollama.
• Prompts to run commands directly (toggleable).

Installation

This package will be installable via pip:

pip install cl-ai

Previously, to run directly from source:

# Make the script executable (if running from source)
# chmod +x cmd_ai.py

# Optional: create a symlink in your PATH (if running from source)
# ln -s $(pwd)/cmd_ai.py ~/.local/bin/cl-ai

Usage

Once installed via pip, you can run cl-ai directly.

Interactive Mode

Run the tool without arguments to enter interactive mode:

cl-ai

You'll get a prompt where you can type your queries:

cl-ai> How do I find large files in the current directory?

After receiving a response, the tool will ask if you want to run the command:

Run this command? [y/n] find . -type f -size +10M -exec ls -lh {} \;:

You can enable or disable this prompt in the configuration.

One-Shot Mode

You can also use it directly from your terminal for a single query:

cl-ai "How do I find files modified in the last 24 hours?"

Commands

In interactive mode, you can use these special commands:

• !help - Show help information
• !config - Configure CL AI
• !history - Show command history
• !clear - Clear command history
• !quit - Exit CL AI

Command-Line Arguments

# Configure the tool
cl-ai --config

# Show command history
cl-ai --history

# Get help
cl-ai --help

Configuration

Configuration is stored in ~/.cl_ai_config.json and can be edited directly or using the !config command in interactive mode or --config on the command line.

Configuration options:

• model: The LLM model to use (e.g., "phi4:latest")
• url: The API URL
• type: The API type (currently only "ollama")
• prompt_prefix: System prompt for the LLM
• auto_run_prompt: Whether to prompt to run commands (true/false)

Running Tests

Unit tests are included to ensure functionality. To run the tests, navigate to the project directory and use the following command:

python -m unittest discover -v

This will automatically discover and run all tests in the test_cmd_ai.py file. (Note: Test file name might need update if module name cmd_ai.py changes).

Examples

# Get a command to find large files
cl-ai> How do I find files larger than 100MB?

# Get a command to monitor system resources
cl-ai> How can I monitor CPU and memory usage?

# One-shot mode
cl-ai "How do I search for text in multiple files?"